back to conference papers page

Conference Papers Online

Keynote Address

Archival Description from a Distant View

Peter Horsman

Netherlands Archives School
International Council on Archives Committee on Information Technology, chair


This paper is about looking at archives and records [1] from a distant position; it is about abstraction by modelling, and about founding archival description on abstract models. Models, which can be used for more than designing archival description alone, but also for better understanding appraisal and other archival methods. It is about the application of archival standards, and eventually about seeking global archival concepts - concepts that archivists all over the world should share, rather than stressing national particularities.

Consequently, this paper might be somewhat abstract, the reader be warned.

I will try to bring together some of the current ideas around archival description, including those of contents, form, context, structure and meta data. I’ll speak about archival constructs as well. Most of these concepts one finds in the conference program, and likely other papers explain these issues better than I can do. I consider my task to stay at a global level, that’s where I feel most comfortable.

I will build my paper upon the generic data model of the archival fonds and its context on which the ICA Committee on Information Technology is working. I will not get into the details of the data model, but rather focus on the archival theoretical concepts underlying it. The model itself was intended to support ISAD compliant database design, and this links my paper to archival standards.

Archival Description

What, then, is archival description, or rather, what are the objects we describe - the sources of the knowledge this conference is about - and that archivists want to control intellectually and administratively. Let us have a look first at the real world of records.

What is it that archivists describe: single sheets of papers, such as letters, or single pages in a volume? Or rather volumes, bundles or folders as a whole? Or is it a series of volumes? Or an entire fonds. Or do we describe the context of the fonds and its records, such as the record creating organisation and its functions and competencies. Or do we even include the context of the records creator into the real world to be caught into our descriptions.

The Problem

In 1993 in an article in Archivaria David Bearman criticised the little progress archivists had made by that time in developing automated systems for archival description. "Indeed", he said, "the history of archival automation has not been a story of great successes." Archivists had not been able to adequately use the tools of information science. [2]

I am not too sure that since Bearman wrote this, really much progress has been made. Certainly, since 1993 much work has been undertaken to establish archival standards for description, such as at the international level ISAD(G) and ISAAR(CPF), or at national levels may be mentioned the Canadian Rules for Archival Description. Concerning computer applications, all over the world archivists are breathless trying to adopt new technologies for controlling records, inventing new techniques, such as Encoded Archival Description. But how innovative are these initiatives?  Do they really bring archival practice forward into the information age? Or is it just old wine in new barrels?

Let us put ourselves the question: [3] aren’t archivists trying to catch the tiger by the tail, in stead of riding the beast? It is, like in 1993, looking at documentary standards, borrowing techniques from librarians, rather than looking at recordkeeping systems and records creating processes. It is still the conceptual thinking of MARC, and still far from archival description for the new millennium.

The Kernel

Let me try to make these points clear, where its all about.

I think that the origins of the relative failure of archival automation, so far, lies in the fact that archivists take too much a documentary approach. Their starting point is the records materials - sometimes referred to with the metaphor ‘by-products’ of business processes; or even worse - ‘sediments of activities’, like polluted mud in an industrial harbour, the dirt of the bureaucracy - good enough only for archivists in the basements of the administration! How many of us do not have at least once the experience in finding themselves digging in piles of neglected papers, in the midst of rusty typewriters, broken chairs, and other remainders of office works (if not worse - I’ve even seen a dead dog, which starved on the files in a local archives). What business are we in, to quote David B. Gracy II, [4] are we in the business of refuse collecting? Records are not the sediment of activities, or the by-product of processes. They are the very atoms of the activities in which they fulfilled their indispensable roles. Often, they are even the only evidence left of the activities; the products, half-products and building blocks of administrative processes. By nature, they are as dynamic as the processes that created and used them. A sedimentary view pushes archives towards a documentary approach, a static view instead of a dynamic perspective - the approach of preserving or reconstructing an assumed old, original order of records, which in reality never existed, because it was as dynamic as the organisation which created and used the records.[5]

A documentary perspective, a static view, brings us to basically documentary methods and techniques, focusing rather on the documents, the records, than on the context in which they have been created. That is what we see happen in applying ISAD standards as a method: starting with the fonds as a dead collection of used papers - and not with the record as the dynamic evidence of a business transaction.

The whole construct seems a library approach, not an archival approach, despite the principles and terminology. This is not a critique, but an observation.

What is a Record?

An archival fonds, or any kind of grouping of records, is fundamentally different from an artificial collection of documents. This is nothing new in archival theory - but why do archivists neglect these basic principles so often in description?

What we really need is to recognise the dynamic nature of records - their bond with the business processes that brought them into life, and that used them. The primary bond is not the bond between records (by nature documents cannot have an interrelationship), neither between the records and their creator - which states the fonds concept - but between the records and the processes, or business transactions. All other relationships, bonds, depend on this one, and do not exist without it.

The primary provenance is a functional provenance. A record is a record, not by nature, but because of its function, its role in one or more business processes. That is the basic idea of the functional archival science, as Angelika Menne-Haritz termed it at the ICA Congress in Beijing.[6]

The business transaction defines the record’s contents and form. The content is the representation, a picture, of the act or fact. The form is dictated by the administrative procedures and legislation applicable to the transaction.

A simple example may clarify the relationships between act, contents, and form. 

In many countries notaries are involved in real estate transactions. When one person sells a real estate property to another person (fact), the notary makes up a legal document, which contains a description of the transaction (contents) in a special, legally defined form. Due to the form (and to the fact that a notary made up the document), the document becomes a legal evidence of the transaction.[7] 

The relationship between record and transaction, illustrated by a simple graph, is the essence of archival science.

click on image to see full-scale version

Two small boxes, and a line, a potentially more powerful model than the hierarchical diagram of the fonds and its component parts, which underlies ISAD. That top-down model is too static, too rigid an archival construct, which does not reflect the reality of record creation. Records, created in a series of organically interrelated business processes, form a fonds; a fonds is to be defined bottom-up, rather than top-down.

The Two Worlds

Indeed, both components, the transaction and the record, are part of a bigger whole. The transaction is part of a complex system of processes, functions and organisational units. The record is part of an often complex system of files, and series, and sub-series - all archival constructs defined by the record keeping system. Both worlds have their own characteristics, rules, behaviour and dynamics.

click on image to see full-scale version

If one wants to describe archival materials, in order to control them, or for whatever purpose, we should look at both worlds separately, but we must understand their relationships, and what governs their behaviour. Describing for access, or intellectual control, the archivist must explain to their potential users the context in which the records live or lived, since the context is crucial for understanding the records.

Record creation has to do with the functions, tasks of the organisation. Recordkeeping takes care of the orderly availability and continuing reliability of the records, by bringing them together into the Records System.[8]


First the left hand part of the model - the world of the business transactions. Not surprisingly, this part is underdeveloped in the current work on archival standards.

Some works has been done in the ISAAR standards, but rather as being sub-alternative to ISAD, than as a separate component of an archival description system. The standards do not support terminology on competencies, functions, tasks and transactions. ISAAR lacks an explicit model of records creation, such as the underlying model for ISAD.

click on image to see full-scale version

The fundamental entities in this part of the description model are the business processes, the organisational functions they carry out, and the organisational structure that is responsible for the operations. At least for public bodies the relationships between organisations and their functions are defined by the governing legal system. The model calls this relationship ‘competence’. As one of the few archival institutions the Swiss Federal Archives undertook long lasting research into the competencies of federal agencies, and implemented the findings into a database - the Kompetenz Kartei. A similar approach is undertaken by some other archives, including the National Archives of Canada, and the Netherlands in the course of developing new appraisal strategies. [9]

In present archival theory this part of the real world of records is referred to as the context - to be precise the context of record creating and the records system.

Recordkeeping System

Before going to the right hand side of the model - the recordkeeping system - I‘d like to stand still a moment at the moment of record creation.

click on image to see full-scale version

First, the momentum of record creation itself, the relationships between process and records are more complex than just making a document. Relationships include also use and deletion of records. Of course, records can only be created and deleted once, but they can be used an unlimited number of times, by various business processes. A record created in an operation might be used for control purposes in another operation.

These are the relationships which David Bearman states should be documented as well. It is impossible to do this afterwards, let us say after records disposition. It should be done by the recordkeeping system, at the very moment of both creation, use and removal. In fact, any description afterwards is due to a defunct recordkeeping system at the moment of record creation and use. One could take the position that archivists, instead of designing description systems, should put their energies in developing better recordkeeping systems! This is particularly true and promising in our present day, in which records are created in an electronic environment and preserved in a digital format. Consequently recordkeeping systems need to be redesigned, and should have the functionality of being ‘self-documenting’.

This is really going into the forefront of records creation, archivists being involved at the conception stage.

Such an ideal type of recordkeeping system will then automatically capture the relevant contextual information, such as the circumstances of creation and use, and capture the data about the technology used at the momentum of creation and use. The latter I’d like to call meta data. The distinction between context information and meta data has been advocated by the ICA Committee on Electronic Records, in its Guide for Managing Electronic Records from an Archival Perspective (1996).

The Records System

Then, finally, the records system, the whole of archival constructs that we call the structure.

ISAD defines a simple, straightforward, hierarchical structure, with the fonds at the top, eventually with sub-fonds underneath, and subsequently series, sub-series, and so on. At the bottom line ISAD puts the item, the smallest physical unit to be produced in the reading rooms. An item could be a single document, a volume or a file.

ISAD does not mention record, but eventually defines the piece as the smallest, logically inseparable unit, such as a single document.

As a matter of fact, one may question whether ISAD is very compliant with current archival thinking. ISAD uses the word fonds, but what does it mean other than a Record Group. By no means ISAD seams to refer to the ideas and concepts as laid down in the publications by Michel Duchein and Terry Cook. [10]

In line with Cook, I would define a fonds as a conceptual whole, consisting in all the records created by an entity (organisation, agency, family, person). It is a concept with, I think, no direct practical meaning. In relational database terminology it is a view, a result of a database query: ‘select all records created/used by ...’

click on image to see full-scale version

The records might be put by the recordkeeping system (or later by an archival institution) into units, such as folders, or volumes. A number of items we might term ‘series’, or ‘class’, or ‘record group’, but what is the essential difference? All archival constructs - what’s in a name?

A flexible concept, as pointed out before, theoretically underpinned by authors such as Terry Cook, must be the basis of archival description, making use of emerging technologies: a data base of descriptions of records, records structures constructed by the recordkeeping system, business transactions structures constructed by the creating entity, and the relationships between them. Archival constructs, such as series, can be created by data base queries, using the criteria for such constructs.

Abandoning the Series System?

In conclusion, we should try to look into our electronic future. Do we need those archival constructs anymore, what do series, or record groups mean in the world of electronic records?

A Recordkeeping System should mirror the business functions and transactions that created and used the records. A file, a case file for example, is the whole of records created and used in the course of one particular set of interrelated business transactions, a case. The idea of case files can be implemented physically by means of folders, but not necessarily. Electronic case files use the same metaphor, but physically the records can be stored in any way. One record can easily belong to more than one case file, even without copying, but by means of  ‘pointers’, or queries.

A series could be a set of files, or just a set of records, created and used in the course of a business function, or following another criterion for establishing a series. Similarly to the case file, a record may belong to more than one series at one moment or over time. An administration is a dynamic environment, records follow administrative changes. As for files, series can be constructed by database views, by queries.

The emergence of electronic records requires rethinking of many concepts in archival science. Does, for example, archival science need the concept of the series - other than just as a term? Is not the series just one archival construct developed for a paper world as a tool to establish administrative and intellectual control at one time?

When Peter Scott in 1966 wrote his famous article on abandoning the record group, didn’t he put a time-bomb under his own series concept? A time-bomb that we will see bursting in the world of electronic records?

Indeed, it is about time not only to redesign archival description, but the whole of our archival methodologies and archival constructs. No doubt, new constructs will be built, to be abandoned in 10, 20 or 50 years. Who cares? As records are as dynamic as the business processes that give rise to them, archival science must be as dynamic as the records and record groupings that are the objects of its research.

[1] Some archival traditions, including Anglo-Saxon and German, distinguish archives from records. The first are records worth preserving for an unlimited period of time (in German: Archivgut and Registraturgut). Other archival traditions, including those of the Netherlands and Italy, do not make such a distinction.
[2] David Bearman, "Record-keeping systems", Archivaria 36 (1993), pp. 16-36.
[3] I borrow the following metaphor from Tjorn-Börn Kjölstadt, Proceedings Expert’s Meeting on Electronic Records, The Hague June 1997, The Hague, 1997.
[4] David B. Gracy II, "Hurtling towards our Jupiter", Janus 1993.2.
[5] Peter Horsman, "Taming the Elephant: An Orthodox Approach to the Principle of Provenance", Kerstin Abukhanfusa and Jan Sydbeck (ed), The Principle of Provenance. Report from the First Stockholm Conference on Archival Theory 1993, (Stockholm, 1994).
[6] Menne-Haritz drew upon a previous article by Bruno Delmas, "Bilancio e prospettive dell’archivistica francese alle soglie del terzo millennio", Oddo Bucci (ed), L’archivistica alle soglie del 2000, Macerata, 1992, pp. 79-107. Also published in French in Gazette des Archives.
[7] The University of Pittsburgh project on electronic records stresses the relationships between business transactions and records, and takes the transaction as a starting point, whereas the University of British Columbia project starts with the record.
[8] Often literature on recordkeeping does not make a clear distinction between the Recordkeeping System and the Records System. For me, a Recordkeeping System is the whole of people, methods, tools, computers, software, procedure, knowledge and records, which are responsible for an organisation’s recordkeeping. The Records System is the (structured) whole of the records created by an organisation and kept by its recordkeeping system. A Records System is thus a sub-system of the Recordkeeping System.
[9] See Helen Morgan, "How do I appraise thee? Let me count the ways: the archival imperative and the construct of appraisal", Working With Knowledge International Archives Conference, 6 - 7 May 1998, Canberra, The USA National Archives and Records Administration currently maintains a Government Locator Index.
[10] T.Cook, "The Concept of the Archival Fonds in the Post-Custodial Era: Theory, Problems and Solutions", Archivaria 35 (1993), pp. 29-37.

back to conference papers page

Published by: Australian Science Archives Project on ASAPWeb
Comments or questions to: ASAPWeb (
Prepared by: Helen Morgan
Graphics by Lisa Cianci
Date modified: 7 October 1999